Livestream longevity, game worlds, Blobperas, and augmented artwork
January 8, 2021 Edition (#37)
Happy 2021, gang. To quote the legendary bards known as House of Pain: “Just like the prodigal son, I’ve returned.” I hope your holidays were healthy, festive, and replete with the illusion that early 2021 would be vastly different from late 2020 despite all earthly evidence to the contrary. Turns out it’s pretty much exactly as awful as it was a week ago, but I’m trying not to languish [Ed.: publicly] in Bostonian cynicism this early in January.
“Don’t worry, everything is going to be O.K.”
Consider the l(ivestream)
This past July, David Peisner, of the New York Times, wrote an article documenting the ubiquity of pandemic era livestreams, and their essential role as a bridge between performers and audiences as we awaited a return to live shows. Nearly six months later, we’re still waiting, and livestreams are very much a part of the fabric of life for creators and their fans. While Peisner’s piece focused on what I’ll call “conventional” livestreams—wherein artists essentially gave their standard live performances, perhaps stripped down, via webcam for a remote audience—Trillium has generally been more intrigued by the “alt-concerts” that we’ve seen occur within various video games.
It’s for this reason that we’ve spent a considerable amount of ink in these hallowed pages documenting various in-game performances. I’m generally not a believer in the longevity of conventional livestreams, once/if the pandemic ends and shows are possible once again. Having watched several performances in this vein over the past 9 months, I can say that I, personally, don’t feel they’re anywhere near a substitute for “the real thing” and, thus, unlikely to maintain their current prevalence. They’re also not that much fun for performers.
I am a believer that alt-concerts, whether in video games or other digital platforms that have large, existing userbases, will persist beyond the pandemic, and even grow in popularity as a new generation of young concertgoers see the experience as something fundamentally different—and perhaps better—than the traditional live show. This is why I’ve been intrigued by startups like Wave, which seeks to make it easier for any artist to put on this type of virtual performance.
It’s also why I’ve been intrigued by the Roblox IPO, formerly intended to occur in December 2020, and whether the markets would see Roblox as “just a game”—albeit a massively popular one—or as a platform for a range of creative experiences. When December came, Roblox reconsidered its IPO plans (NYT), seemingly owing to the frothiness of the markets and questions around how to avoid leaving money on the table. On January 6, 2021, Roblox announced that it had instead decided to go public through a direct listing (VentureBeat), and had simultaneously raised a fresh $520M at a $29.5bn valuation. If you’re keeping score, that’s a 7x increase in valuation since Roblox’s last funding round in February 2020. Pretty standard stuff, really.
The news also occasioned an incredible, Silicon Valley-style quote from Roblox’s CEO, in which he described Roblox’s mission as “[building] a human co-experience platform that enables shared experience, from play to work.” [Ed.: we just wish they dreamed bigger!] The company saw ~160M installs in 2020 (+43% YoY) and $1bn+ in mobile in-game consumer spending, so perhaps a tad of grandiosity is justified. Most interestingly, from our perspective, 960,000 developers earned Robux by producing apps for and within the game, thus justifying Roblox’s claim to be a platform for a wider variety of experiences. [Ed.: There’s no comparison between the companies, but I wish we weren’t reminded of WeWork CEO Adam Neumann’s claim that his real estate company’s mission was to “elevate the world’s consciousness.”]
To bring us back to where we started, with Peisner’s livestream article, it’s worth a scan of this piece in IQ, in which Roblox’s head of music talks about his view of the nascent virtual concert market. I generally find the thesis that people, especially in Gen Z, are turning to Roblox for “social connection,” and that the type of experiences facilitating that connection will naturally evolve, to be quite credible. With 36M+ people around the world already visiting Roblox daily, there’s no reason why it—and other similar “game worlds”—couldn’t replace the incumbent social networks of today as they become the dominant online communities for myriad types of interactions.
OnlyCameos
In addition to catalyzing the livestream revolution, Covid-19 accelerated many companies that are building products to help creators get paid. This week, The Information’s Alex Heath wrote a great (but paywalled) article about Cameo: “a service that lets fans pay celebrities to create personalized video messages.” Heath described how the startup, currently valued at $300M, is thinking about expanding its offering to encompass new monetization types (like subscription) and product features (in-app video chat).
What was most interesting to me, though, was the extent to which Cameo is replicating the features offered by other creator-focused companies like Patreon (paid memberships) and OnlyFans (paid exclusives and direct messaging). There is, to some degree, differentiation in terms of the creators that these companies serve; Patreon targets longer-tail creators, often YouTubers, while OnlyFans started with sex-related content, and Cameo owns the more traditional celebrity angle. As these companies converge, though, I wonder whether we will see any sort of consolidation, or differentiation on a geographic basis. The latter seems to be beginning, with Patreon seeming to focus on EU expansion, while Cameo sees promise in India and Latin America.
There’s also an intriguing trend around business-focused offerings within these companies. Cameo for Business is one such initiative, connecting brands and “influencer” talent through a two-sided marketplace where transactions are brokered through Cameo’s platform. In a different but thematically related direction, OnlyFans just announced that Vice became “the first publisher” to launch a verified account, specifically for its Munchies brand. (Tubefilter) This is significant both because it establishes OnlyFans as a viable home for non-sex-related content, and as a sensible direct-to-consumer commercial extension for media companies, not just individuals. [Ed.: thanks to Down-Under Reader JW for the tip.]
One thing is for certain: as more and more creatives struggle economically due to the pandemic’s impact (NYT), there will be a growing demand for products that can help them replace income, and an ongoing effort amongst tech companies big and small to build products that meet this need.
I wish someone would replace me with an AI
But, until that magical day comes, it’s business as usual. The same might not always be true for human musicians, according to a small but increasingly vocal group of music industry analysts and writers. For some months, various sources have speculated about Spotify’s long-term intent to “replace” human artists with AI composers.
The general logic is the well-documented tension between the company and its musician-stakeholders, centered on complaints from the latter regarding paltry per-stream royalties and their perception of an unfair value exchange. Spotify, they claim, reaps the benefits of millions of consumer users—via ads & subscription—but doesn’t appropriately compensate the original producers of the products that draw those users to the platform. Given this dynamic, it’s not too much of a stretch to argue that Spotify might think it strategic (mL) to develop alternative sources of inventory that have less complicated commercial implications. Indeed, the company’s investment in non-music audio—specifically podcasting, through acquisitions of Gimlet, The Ringer, Anchor, and others, as well as original programming—can be seen as proof of this strategy in action.
It’s a bit more of a leap, shall we say, to believe that Spotify truly thinks that it could use AI-generated music as a meaningful portion of its available library. Sure, maybe it’s not as challenging for some instrumental, ambient, or beat-driven genres; but, most consumers are probably going to want their Beyonce and Beatles, and not the creepy AI versions thereof.
Nonetheless, Input’s J. Fergus recently wrote a piece where he referenced a new Spotify patent, also referenced by Music Business Worldwide, that theoretically allows the company to “police songwriter plagiarism.” On paper, the patent covers a technology that would compare a song’s melody, chord structure, and other musical elements to all other songs within Spotify’s library, thereby calculating a “similarity value” between that piece and its peers. From a different angle, Fergus argues, the technology could be used to generate “new lead sheets” or musical structures, thus resulting in AI-created artifacts.
One of the inventors credited on the patent is François Pachet, director of Spotify’s Creator Technology Research Lab, and the guy who’s responsible for designing the “next generation of AI-based tools for musicians.” One could argue that this tool does indeed help musicians by allowing them to quickly identify plagiarism of their work over a massive dataset, with downstream implications for royalty claims and counter-piracy. Or maybe it’s all a sinister conspiracy and mankind is doomed.
I personally don’t lose sleep over the replacement of human artists with AIs—[Ed.: there are plenty of other things I can recommend if you want to lose sleep over something]—and am more compelled by a vision of hybrid human/AI creations where the latter increases the creative reach of the former. That said, it’s an admittedly slippery slope, and one could imagine a world in which automated creations form a sort of next-gen elevator muzak, thus further endangering the livelihoods of human artists.
This is to say nothing of the generally hilarious sub-genre of synthetic art in which well-known human artists are deepfaked into performing novel works. See Jay-Z rapping Shakespeare:
Jay was not amused (yr.media), but I am. Obviously, this could get dark in a hurry, in the wrong hands; it’s fun to hear Obama reciting Biggie’s “Juicy” but would be somewhat less fun to hear him lend his dulcet tones to one of Trump’s barn-burners. I’m sympathetic to an artist’s right to own their voice, the same way they own how their physical presence is represented, and see this as tied to the complicated ethical and representational questions at play when dead artists are reincarnated to perform in hologram form à la Tupac. How we regulate and enforce this ownership on an increasingly vast internet, as these technologies become more distributed and sophisticated, is a question for another day.
Instead, let’s revel in the joys provided by a less fraught computer creation: the “Blob Opera,” a machine learning-based experiment originated in Google’s Arts and Culture Initiative. This is honestly the best thing that’s happened to me in decades [Ed.: actually] and I highly encourage all of you to check out the absolutely spectacular, Grammy nominated, Pavarotti-approved ~1min long composition that I created as my holiday gift to the the Trilliuminati. As Chappelle once said, turn your headphones up.
Robogirlfriends, augmented sculptures, and LiDAR effects
In my end-of-year search for articles about ways that technology is entering new areas of the human experience—[Ed.: in between futile attempts to reseed the Trillium HQ lawn], I discovered a fairly staggering number of articles, most of which I’ve chosen to burn upon reading (@Coens). Three quick hits to follow.
First, it’s worth reading Zhang Wanqing’s article (Sixth Tone) on “the AI girlfriend seducing China’s lonely men.” It’s exactly what it sounds like! A “sassy bot” named Xiaoice, driven by AI, that communicates with “millions” of users: 660M around the world, 3/4 of whom are men. What could go wrong? Originally developed by researchers inside Microsoft Asia-Pacific in 2014, the bot has evolved visually and in terms of its “empathic computing framework.” While the idea of chatbots providing emotional support is gaining mainstream acceptance, and may be healthy in specific therapeutic interactions, there’s an obvious danger to humans becoming dependent on AI creations for companionship and emotional fulfillment. There’s also risk implicit in the vast amount of intimate data that is being created and captured over the course of these conversations.
Second, I enjoyed this writeup on Acute Art and Dazed Media’s “Unreal City” activation, which was “London’s biggest public festival of AR art” and featured 36 sculptures arranged along the River Thames. As you can see below, the exhibition revolves around the idea of multiple artists creating AR works that are accessible to people using a special app and visiting the 24 sites across which the activation unfolds. At each location, the AR art “materializes” and is visible as an overlay on the physical world, meaning that otherwise familiar public spaces are transformed into a sort of outdoor art gallery. Check out the video synopsis below.
Third, and along similar lines, TikTok recently rolled out its first “lidar-powered AR effect.” (Tech Crunch) The effect centered on New Year’s Eve and showcased an “AR ball,” filled with confetti that would drop and explode after a short countdown. This experiment is interesting to me largely for the same reasons that I find Snap’s focus on AR lenses compelling; because of the large user communities that these apps have cultivated, many consumers will be exposed to augmented reality through the portal of TikTok or Snap. The quality of the immersive activations that these companies enable, and how they strike “the average individual,” will play a significant role in determining the hype and momentum surrounding mass-market AR technologies in the coming years. Also, all due respect to TikTokComms, but I have never been this excited about New Year’s.
For your ears only
One reliable source of peace in troubled times is the combination of Anderson Paak’s voice and Knxwledge’s production. The duo, jointly known as NxWorries, released one of my favorite albums in 2018—Yes Lawd!—and have continued to intermittently drop excellent new music. Most recently, they performed a new track called “Where I Go,” which exists in various bootleg formats across the internet (including below). For good measure—and due to uncanny thematic, aural, and visual coherence—I’ve included Tom Misch & Yussef Dayes’ “Nightrider,” featuring the always impressive Freddie Gibbs. These are convertibles; no masks required.
See you all next week.
N